OVERVIEW OF CONVERSION PROCESSES...

Transcription

1 October 2005 Gasification-Based Fuels and Electricity Production from Biomass, without and with Carbon Capture and Storage Eric D. Larson,* Haiming Jin,** Fuat E. Celik* * Princeton Environmental Institute, Princeton University, Princeton, NJ **Thayer School of Engineering, Dartmouth College, Hanover, NH 1 INTRODUCTION OVERVIEW OF CONVERSION PROCESSES STATUS OF TECHNOLOGIES PRODUCTION OF CLEAN SYNTHESIS GAS Biomass Gasification Gas Cleanup CONVERSION OF SYNTHESIS GAS TO FINAL PRODUCTS Synthesis of Fischer-Tropsch Fuels Synthesis of Dimethyl Ether Hydrogen Production DETAILED PROCESS DESCRIPTIONS AND MASS/ENERGY BALANCES PRODUCING CLEAN SYNTHESIS GAS STAND-ALONE ELECTRICITY PRODUCTION Electricity Production with No Carbon Capture and Storage Electricity Production with Carbon Capture and Storage DME PRODUCTION DME Production with No Carbon Capture and Storage DME Production with Carbon Capture and Storage FISCHER-TROPSCH FUELS PRODUCTION F-T Fuels Production with No Carbon Capture and Storage F-T Fuels Production with Carbon Capture and Storage HYDROGEN Hydrogen Production with No Carbon Capture and Storage Hydrogen Production with Carbon Capture and Storage COST ANALYSIS DEFINITIONS AND METHODOLOGY CAPITAL AND PRODUCTION COST ESTIMATES Stand-Alone Electricity Production Fuels Production ACKNOWLEDGEMENTS REFERENCES... 72

2 1 Introduction We report here on design, mass-and-energy-balance calculations, and production cost estimates for gasification-based thermochemical conversion of switchgrass into Fischer-Tropsch (F-T) fuels, dimethyl ether (DME), and hydrogen, in all cases with some level of co-production of electricity. Also, some process designs are developed and analyzed that include capture of byproduct CO 2 for underground storage. Additionally, we present results for stand-alone electricity production using integrated gasification combined-cycle technology, both with and without carbon capture and storage (CCS). The feedstock considered in all cases is switchgrass, and the reference production scale is an asreceived input of 5,670 metric tons per day (tpd) of switchgrass having a moisture content of 20%. 1 This corresponds to a dry matter flow of 4,536 tpd (or 5000 dry short tons per day). The 20% moisture level is sufficiently low that active drying of the feed material is not necessary before gasification. This saves considerable capital cost by avoiding a dryer, while imposing little if any efficiency penalty relative to systems with active drying utilizing low-grade waste heat. The physical and chemical characteristics of the assumed switchgrass are given in Table 1. The energy flow corresponding to 5,670 tpd of 20% moisture switchgrass is 983 MW higher heating value or 893 MW lower heating value. Table 1. Characteristics of switchgrass assumed in this analysis. a As-Received Proximate Analysis Fixed carbon (wt%) 17.1 Volatile matter (wt%) 58.4 Ash (wt%) 4.6 Moisture content (wt%) 20.0 Lower Heating Value (MJ/kg) 13.5 Higher Heating Value (MJ/kg) 15.0 Ultimate Analysis (dry basis) Carbon (wt%) 47.0 Hydrogen (wt%) 5.3 Oxygen (wt%) 41.4 Nitrogen (wt%) 0.5 Sulfur (wt%) 0.1 Ash (wt%) 5.7 Lower Heating Value (MJ/kg) 17.5 Higher Heating Value (MJ/kg) 18.7 (a) The proximate and ultimate analyses are for whole-plant Alamo Switchgrass as given in the US Department of Energy database available at (entry #74). We have normalized the original DOE ultimate analysis (carbon, 47.27%; oxygen, 41.59%; hydrogen, 5.31%; nitrogen, 0.51%; sulfur, 0.1%; and ash, 5.76%), so that the percentages sum to 100%. The higher heating value given in this DOE database is 7998 Btu/lb (dry basis), or MJ/kg. We used the latter value as an input to our Aspen process simulation. Aspen uses the initial value in calculations involving values of other thermodynamic properties. The latter values vary slightly depending on which one of Aspen s internal thermodynamic property sets is selected for use in the simulation. As a result, back-calculating from the final Aspen simulation results yields a slightly different higher heating value (shown in the table), but one that is consistent with the overall Aspen calculations. 1 The average as-received moisture content for large-scale production of field-dried switchgrass is expected to be 15% to 20% by weight (S. McGlaughlin, personal communication, October 2003). 2

3 For all process designs considered, we have developed energy and mass balances (using Aspen- Plus process simulation software) for N th plant systems, i.e., systems for which it is assumed that all research, development, and demonstration hurdles to commercial implementation have been overcome. This requires projecting some performance characteristics and costs, since none of the systems under consideration have reached commercially mature performance, reliability, or cost levels today. There have been considerable technology developments in closely related areas, as discussed in Section 3, which provides a good basis for making N th plant performance and cost projections. As discussed by Larson (2003) and Larson, et al. (2004), the most important technical hurdles that have yet to be overcome, but which are assumed to have been overcome for the purposes of this analysis, relate to a) reliable feeding of low bulk-density biomass like switchgrass into a pressurized gasifier vessel without excessive feeding-energy requirements, b) high-reliability operation of commercial oxygen-blown fluidized-bed gasification, c) essentially complete cracking of tars in the raw gasifier product into non-condensible combustible gases, d) raw gas cleanup of particulates, alkali, sulfur, and other trace contaminants to the specifications needed for downstream processing, and d) tight process heat integration and control, resulting in maximum recovery and use of process waste heat. Based on the required equipment capacities derived from our process simulations, we have developed capital cost estimates by major plant area drawing on literature sources, extensive discussions with industry experts, and our own prior work. The uncertainty in our capital cost estimates is ±30%, based on the level of detail (factored estimates for major plant areas), the nature of literature and industry sources (including some equipment-vendor quotes) from which we derived costs, and the inherent uncertainties in projecting N th plant costs given the precommercial status of some of the major pieces of equipment included in the systems examined. Using widely-accepted equipment-level cost scaling exponents, we also estimate capital costs for plants of different capacities from the reference capacity to illustrate scale economies inherent in the technologies. It may be noted that the reference biomass feed rate (5,670 tpd) is considerably larger than has been considered previously for biomass conversion facilities in most analyses, although a few commercial lignocellulosic-biomass processing facilities this size are in operation today. 2 The primary incentive for building plants this large is more favorable economics. While it is feasible and desirable to build large-scale biomass conversion facilities, most analysis and commercial implementation of bioenergy to date has focused on relatively small-scale conversion plants. This is due largely to the prevailing thinking that low-cost biomass feedstocks (residues and wastes) are necessary for the viability of projects in the near term. Residues and wastes are relatively dispersed resources, so large quantities cannot be cost-effectively brought to single sites, thereby limiting the size of conversion facilities. With dedicated production of switchgrass for energy, it becomes plausible for large-scale conversion facilities to be supplied with feedstock from within reasonable distances (incurring 2 For example, in Brazil there are an estimated 15 sugarcane processing facilities that consume sugarcane at a rate of 5,000 dry metric tons/day or greater (personal communication from J.H. Suleiman, Centro de Tecnologia Copersucar, Piracicaba, Sao Paulo, Brazil, 6 May 2004). Since typical Brazilian sugarcane has a moisture content of about 70% when it reaches a mill, 5000 dry tpd corresponds to about 17,000 as-received tpd. 3

4 only modest transport costs), particularly in areas where high switchgrass yields can be achieved. Figure 1 shows the approximate transport distances that would be involved in supply 5000 metric tonnes per day (dry matter) to a central conversion facility as a function of yield and planting density. Shown for reference are planting densities for corn across corn-growing regions of the United States, as well as the planting density for corn in the heavy corn-producing region of Southeastern Iowa. Radius of circular area, km Radius for 5000 dry t/day biomass supply (~1000 MW b ) Yield on planted area (dt/ha/yr) Typical U.S. corn density Southeast Iowa corn density 0 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Percent of circular area planted Figure 1. Calculated radius of circular area from which switchgrass would be harvested for delivery to a conversion site located at the center of the collection area. Radius is shown for different average yields (dry metric tonnes per hectare per year) and percent of area cultivated. In this report, we begin with an overview of the process designs for converting switchgrass into fuels and/or electric power. We then summarize the commercial status of relevant technologies. Detailed process design and simulation results are then presented, followed by capital and production cost estimates. Some relevant comparisons between the estimated production costs of biomass-derived fuels and power and fossil-fuel based products are also given. 2 Overview of Conversion Processes In a later section of this report we will present detailed results for a large number of different process configurations. To help the reader navigate through the different process designs, we begin here with simplified descriptions. All processes examined here share a common upstream design for converting biomass into a clean synthesis gas. Switchgrass is brought from a short-term storage area into a feed preparation area where it is chopped and conveyed to the gasifier. Gasification is carried out in a pressurized, oxygen-blown, fluidized-bed gasifier, with oxygen provided from an air separation 4

5 unit (ASU). Oxygen gasification is preferred to air-blown gasification because it minimizes downstream equipment sizes (by eliminating nitrogen dilution in process streams) and improves reaction rates in fuels synthesis steps. 3 The gasifier produces a mixture of light combustible species (primarily CO and H 2, with some CH 4 ), heavy hydrocarbons ( tars and oils ) and unwanted minor contaminants (H 2 S, NH 3, HCN, and others). In our simulations, all carbon in the biomass is assumed to be completely converted to gas. The raw gas (at about 1000 o C) is subjected to several cleaning and cooling steps before further processing. These include thermal/catalytic cracking of tars/oils, partial gas cooling (to about 350 o C), and particle removal by filtration (which also removes alkali compounds condensed on the particles). After a clean synthesis gas is produced, the downstream processing of the gas varies depending on what products are being made and on whether or not CO 2 is captured for storage. Table 2 summarizes all cases considered in this work. The case names shown there are used throughout this report. All case names ending in VENT involve no capture of CO 2. All cases with names ending in CAP or CCS involve some extent of carbon capture and storage (CCS). In all cases, electricity is generated in excess of the amount needed to meet onsite process needs. For three of the cases (indicated in Table 2), the amount of excess electricity is relatively modest, for reasons explained below. Table 2. Process designs examined in this work. Abbreviation for Products Carbon Captured? Process Configuration Electricity F-T Fuels DME H 2 NO YES Electricity only BIGCC-VENT BIGCC-CCS Dimethyl Ether D-OT-VENT D-RC-VENT low D-OT-UCAP D-RC-UCAP low D-OT-DCAP Fischer-Tropsch Fuels FT-OT-VENT FT-OT-UCAP Hydrogen HMAX-VENT low H50/50-VENT HMAX-CCS H50/50-CCS Two designs for stand-alone electricity production are developed. Both involve the combustion of the clean synthesis gas in a gas turbine combined cycle. In the case with venting of CO 2 3 Synthesis involves reactions that are driven by the partial pressures of the CO, H 2, and other reacting species in the gas. Inert nitrogen would reduce the partial pressures of the reacting species and lead to lower synthesis rates. An alternative to oxygen-blown gasification for producing a synthesis gas undiluted with nitrogen is gasification driven by heat delivered indirectly to the biomass. Two indirectly-heated gasifier technologies have received considerable development and demonstration support from the US Dept. of Energy during the past 15 years. These include the technology originally developed at the Battelle Columbus Laboratory, in which hot sand delivers heat to a fluidizedbed reactor, and the technology of the company, MTCI, in which heat exchanger tubes are immersed in a fluidized bed. Both of these technologies operate at near-atmospheric pressure, which makes it difficult and/or costly to scale up to the large plant sizes being considered here. 5

6 (BIGCC-VENT), the clean syngas is sent directly to the gas turbine combined cycle (Figure 2, upper). In the case with CO 2 capture (BIGCC-CCS), the clean syngas undergoes water gas shift reaction followed by CO 2 removal, resulting in a hydrogen-rich fuel gas stream going to the GTCC (Figure 2, lower). Biomass Biomass Preparation Pressurized Gasification Gas Cooling & Cleanup GTCC Net Electricity oxygen ASU nitrogen air Biomass ASU oxygen air nitrogen Biomass Preparation Pressurized Gasification Gas Cooling & Cleanup Water Gas Shift CO 2 Removal GTCC Net Electricity Underground Storage CO 2 Figure 2. Simplified process diagram for electricity production from biomass without carbon capture (upper figure) and with carbon capture (lower figure). For production of dimethyl ether, two basic process designs are examined for the case without carbon capture (Figure 3). In both designs, the small amount of H 2 S (~ 500 ppmv) remaining in the clean syngas is removed prior to synthesis to avoid poisoning of the downstream catalyst, and most of the CO 2 in the gas is also removed in order to improve conversion rates in the synthesis reactor. CO and H 2 catalytically combine in the synthesis reactor, which produces a mix of DME and unconverted synthesis gas. In one design (D-OT-VENT, Figure 3, upper), the syngas is passed once through ( OT ) the synthesis reactor. The DME is then separated from the unconverted gas, and the latter is burned in a gas turbine combined cycle to generate electricity, a small portion of which is consumed on site and the balance of which is sold. In the second design (D-RC-VENT, Figure 3, lower), most of the unconverted syngas after synthesis is recycled ( RC ) to the synthesis reactor to produce additional DME. A small purge gas stream is burned in the power island to generate a modest amount of electricity, most of which is consumed on site to meet process needs. For DME production with CCS, three process designs have been developed (Figure 4). With once-through synthesis, CO 2 is captured upstream of the synthesis reactor (D-OT-UCAP, Figure 4, upper) or both upstream and downstream of the reactor (D-OT-DCAP, Figure 4, middle). For the latter case, the unconverted synthesis gas leaving the DME separation area is subjected first to a water gas shift reaction to increase the amount of CO 2 that can be subsequently removed. The third DME process design with CCS is with recycle synthesis (D-RC-UCAP, Figure 4, lower). In this design CO 2 is captured only upstream of the synthesis reactor, since there is relatively little CO 2 available for capture downstream. 6

7 Biomass Preparation biomass Gasification oxygen Air Separation Unit Gas Cooling & Cleanup air H 2 S, CO 2 Removal Synthesis Separation unconverted synthesis gas GTCC process electricity air Liquid Fuel Export Electricity Biomass Preparation biomass Gasification oxygen Air Separation Unit Gas Cooling & Cleanup H 2 S, CO 2 Removal Synthesis process electricity unconverted synthesis gas Separation purge gas GTCC Liquid Fuel Small Export of Electricity Figure 3. Simplified process diagrams for co-production of liquid fuels (DME or FT) and electricity without CCS. The upper configuration shows once-through synthesis, with co-production of a significant amount of electricity from unconverted syngas. The lower configuration (D-RC-VENT) shows recycle of unconverted syngas to synthesis, resulting in less electricity co-production. air Biomass Preparation biomass D-OT-UCAP Gas Cooling Gasification & Cleanup oxygen ASU air (from GT) H 2 S, CO 2 Removal Underground storage Synthesis Separation unconverted synthesis gas process Power Island electricity Liquid Fuel Export Electricity Biomass Preparation biomass Gas Cooling Gasification & Cleanup oxygen ASU air (from GT) H 2 S, CO 2 Removal Synthesis Separation unconverted synthesis gas Water Gas Shift Liquid Fuel Underground storage CO 2 Removal D-OT-DCAP process electricity Power Island Export Electricity Biomass Preparation biomass D-RC-UCAP Gasification oxygen ASU Gas Cooling H 2 S, CO 2 & Cleanup Removal air (from GT) Underground storage unconverted synthesis gas Synthesis Separation purge gas process Power Island electricity Liquid Fuel Small Export Of Electricity Figure 4. Simplified process diagrams for co-production of DME and electricity with carbon capture and storage. The upper and middle configurations involve once-through synthesis, with co-production of a significant amount of electricity from unconverted syngas. The lower configuration involves recycle of most of the unconverted syngas to the synthesis reactor, resulting in a smaller amount of electricity co-production. 7

8 For the production of F-T fuels, we have developed two process designs, both involving oncethrough synthesis. The case without CCS is depicted in the upper diagram in Figure 3 (FT-OT- VENT). The design of the case with CCS, FT-OT-UCAP, is analogous to the DME case, D-OT- UCAP, shown in the upper diagram of Figure 4. We have considered only once-through synthesis for F-T fuels, because the high single-pass conversion of syngas that can be achieved with F-T synthesis (compared with DME synthesis) leaves relatively little unconverted gas for recycle. For hydrogen production, the clean syngas leaving the gasifier/gas cleanup area is subjected to water gas shift, after which CO 2 is removed (for venting or for capture) and hydrogen in the remaining gas is purified. The purge gas from the purification process is burned to generate electricity. Two process designs are considered with CO 2 vented. One case maximizes the production of hydrogen (H-MAX-VENT, Figure 5) by the design of the separation area. The amount of exportable electricity generated in this case is relatively small. In the second case, the separation area is designed so that roughly equal amounts of hydrogen and exportable electricity are produced (H-50/50-VENT, Figure 5). For hydrogen production with CCS, essentially the same designs as with CO 2 venting would be utilized, but rather than venting the CO 2 before the hydrogen purification step, the CO 2 would be captured and compressed for transporting to storage (Figure 5, lower two diagrams). Biomass Gas Cooling Water Gas H 2 S, CO Gasification 2 Preparation & Cleanup Shift Removal biomass H-MAX-VENT Air Separation Unit air Process electricity Separation purge gas Power Island Hydrogen Small Export of Electricity Biomass Gasification Preparation biomass H-50/50-VENT Air Separation Unit Gas Cooling & Cleanup Water Gas H 2 S, CO 2 Shift Removal process electricity Separation purge gas GTCC Hydrogen Export Electricity Biomass Gasification Preparation biomass H-50/50-CCS H-MAX-CCS Air Separation Unit Gas Cooling Water Gas & Cleanup Shift air from GT Underground Storage H 2 S, CO 2 Removal Process electricity Separation purge gas GTCC Hydrogen Export Electricity Biomass Gas Cooling Water Gas H 2 S, CO Gasification 2 Preparation & Cleanup Shift Removal Air Separation Unit air Underground Storage Process electricity H-MAX-CCS H-50/50-CCS biomass Separation purge gas Power Island Hydrogen Small Export of Electricity Figure 5. Simplified process diagrams for co-production of hydrogen and electricity with carbon capture and storage. The upper two configurations are with venting of CO 2. The lower two configurations are with carbon capture and storage. 8

9 3 Status of Technologies Many of the key component technologies in the process designs sketched above are already commercial, while a few key ones are yet to see any commercial use. In this section we discuss the development status of key biomass-specific technologies and also the status of synthesis gas processing technologies that are not specific to any particular starting feedstock. 3.1 Production of Clean Synthesis Gas Biomass Gasification Biomass gasifiers are not commercial for the scale and type of application considered in this analysis, but there is a worldwide transition underway in the fossil fuel sectors toward gasification a transition that is so far confined mainly to petroleum refining/chemical process industries for the co-production of chemicals and electricity [60 GW syngas capacity installed worldwide, growing at about 3 GW per year (Simbeck, 2004)]. The cost-cutting experience being accumulated for large-scale fossil-fuel gasification technologies is likely to spill over to gasification-based conversion of any low-grade carbonaceous feedstock, including biomass, so that once biomass gasification technologies are introduced commercially at large-scale, costs and reliability may approach N th plant levels relatively quickly. While there are no large-scale pressurized biomass gasifiers in commercial operation, development and pilot-plant demonstration efforts with oxygen-blown fluidized bed gasification (the gasification design utilized in our analysis) date to the early-1980s in Sweden (Engstrom et al., 1981; Strom et al., 1984) and the mid-1980s in the USA (Kosowski et al., 1984; Evans et al., 1987). Most such efforts were curtailed when world oil prices fell in the late 1980s. With growing interest in hydrogen as an energy carrier, there has been some recent re-assessment of pressurized oxygen-blown gasification (Lau et al., 2003). Thus, a not-insignificant knowledge base already exists relating to pressurized oxygen-blown gasification technology. The ability to efficiently feed a low-density biomass like switchgrass into a pressurized gasifier is important for achieving good overall process conversion efficiencies and lower costs, especially at larger scales. Pressurized biomass feeding is a technology area where additional development is needed. The conventional approach for feeding low-density biomass such as chopped switchgrass against pressure is through use of lock-hoppers. Conventional lock-hopper technology suffers from high consumption of inert pressurizing gas and the associated gas compression work required. A number of efforts have been made to develop alternative feeder systems to address the drawback of lock-hoppers (Wilen and Rautalin, 1993), including rotary types and plug types (screw, piston, and hybrid screw/piston). None of these have been yet proven viable for large-scale commercial operation. Lau, et al. (2003) provide the most comprehensive recent review and evaluation of alternative feeder systems for pressurized gasifiers. 4 They conclude that for elevated reactor pressures (> 30 bar) lock-hoppers are the only well-proven option. For these pressures, they are pessimistic about the future feasibility of the plug-type feeder designs, especially for fibrous feedstocks such as bagasse and switchgrass. Nevertheless, they recommend further investigation of one of the piston-type feeder designs they evaluated. They propose double lock-hopper and hybrid lock-hopper/plug concepts that might considerably reduce the consumption of inert gas without significant added cost. 4 See Appendix B in Lau et al. (2003). 9

10 3.1.2 Gas Cleanup Maniatis (2001) and others have identified gas cleanup (especially tar removal or destruction) as the most important area where technological advances are needed in order to facilitate widespread commercialization of gasification-based biomass conversion systems. Different authors report different required contaminant removal levels, but by any measure the tolerance to contaminants of downstream fuels synthesis and advanced power generation processes is low Tars Much work has been done on various aspects of tar production and destruction as related to biomass gasification since the 1970s. Several of the more recent and comprehensive review articles on the subject are discussed here. Milne et al. (1998) provide an authoritative review of issues and literature relating to biomassgasifier tars, their production, their measurement and analysis, their tolerance by end-use devices, and their removal or destruction. The article includes a bibliography of some 400 publications that were reviewed in the course of that work. Stevens (2001) defines tars as any organic molecule with molecular weight greater than benzene. Raw gas from a circulating fluidized bed gasifier (CFB) without any primary or secondary treatment of tar, typically has a tar concentration in the range of 1 to 15 g/nm 3. Tar removal (as distinguished from tar cracking) is most commonly done by cooling it to condense the tar into droplets and then removing the droplets much the way particles can be removed (using scrubbers, electrostatic precipitators, etc.). Collected tars can be re-injected into the gasifier, which will cause a portion of them to crack to lighter molecules (a generally desirable result since the energy content of the tars is then largely maintained in the gas). However, it will also cause a portion of the tars to undergo further dehydration and condensation reactions to form highly-aromatic and more-refractory tars (a generally undesirable result). Devi et al. (2003) provide a comprehensive review of tar removal by primary treatments, which he defines as those applied inside the gasifier, as contrasted to treatments of tar-laden gas in external reactors. A positive feature of successful primary treatment is the elimination of the cost and added operating complexity with secondary (post-gasifier) treatments. The main negative feature of primary treatment is the potential detrimental impact on gas quality (e.g., injection of secondary oxidant may lead to a reduced heating value for the fuel gas). Primary treatments include staging of the gasification process, using catalysts, and using injection of secondary oxidant. The authors conclude that for some applications of the product gas, primary treatment approaches that are already known can provide sufficient tar removal. 5 They indicate that the tar removal requirements for any application can be met by a combination of known primary and/or secondary treatments. They indicate that further research is needed before primary treatment alone will be able to meet tar removal requirements for the most stringent applications. Tars can be cracked using heat alone (thermal cracking) at temperatures > 1200 o C, although this is too high a temperature for most biomass applications (wherein ash melting would likely be a 5 For example, Pan et al. (1999) showed in laboratory-scale experiments that injection of secondary air to a fluidized-bed gasifier operating on forestry residues at temperatures above about 840 o C decreased tar content in the raw product gas by 90%, which will be sufficient for a number of applications. 10

11 problem). With assistance of catalysts, cracking can occur at o C. Tar cracking can be carried out in the gasifier bed itself. Difficulties with using catalysts in a gasifier can include physical attrition and catalyst deactivation. Based on tests at a Swedish biomass-igcc demonstration plant in the 1990s (Sydkraft et al., 2001) in situ dolomite catalysts were successful in reducing, though not entirely eliminating tar leaving a pressurized air-blown fluidized-bed gasifier. Using a separate tar cracker, 95-99% of tars in gas streams at o C have been cracked using dolomite under laboratory conditions. A proprietary, disposable, non-metallic catalyst, identified as DN34, developed at Battelle Columbus Laboratory (BCL), has been tested in a fluid-bed tar cracking unit operating at 800 o C with a feed gas from the pilot-scale (10 tonne per day biomass feed) pyrolytic BCL gasifier in Columbus, Ohio. Tar levels in the gas were reduced from 23 g/m 3 to 1.4 g/m 3. Tar conversion was lower at lower temperatures. The syngas was subsequently run through a wet scrubber, and the resulting tar concentration was reported to be essentially zero. Metallic tar cracking catalysts, particularly those based on Ni/Co/Mo blends, also have been shown to be effective at tar cracking. They also destroy ammonia (see below). Tar cracking (with metallic or non-metallic catalysts) in continuously operating, full commercialscale biomass systems has not yet been demonstrated. Finally, Boerrigter and van der Drift (2004) conclude from a careful evaluation and analysis of tar issues around biomass gasification that the optimum technology for large-scale biomass conversion to fuels and chemicals is entrained flow gasification essentially the same gasifier design used commercially for modern coal gasification. The high gasification temperatures with the technology ensure complete tar cracking in the gasifier. The authors acknowledge other technology challenges that entrained gasification would bring (feed preparation and feeding, slag handling, etc.), but are more optimistic about the prospects for finding solutions to these issues than solving the problems associated with tar cracking with lower-temperature biomass gasifiers. Interestingly, the Shell Company, which offers one of the leading commercial entrained-flow coal gasifiers, recently announced a partnership with Choren, a German company developing a biomass gasification system for liquid fuels production based on entrained-flow gasification (Shell, 2005) Other Gas Contaminants Stevens (2001) provides the most recent comprehensive review of the status and future prospects for conditioning of gasified biomass for a range of applications, from those that are relatively tolerant of contaminants in the gas to those that have very stringent gas quality requirements. In addition to tar, his review addresses particulates, alkali compounds, ammonia, and sulfur. Many of the gas conditioning technologies that he mentions are well established commercially in applications other than with biomass gasification. Other of the technologies have been shown to work in small-scale or short-duration tests with gasified biomass, but require scaled-up longerterm testing before they can be considered proven. Removal of alkali vapors is required upstream of devices on which the vapors would otherwise condense due to low temperatures, and thereby cause deposition and corrosion problems. Condensation of alkali vapors typically begins around 650 o C. At temperatures below this, the vapors will condense on particles in the gas stream or simply condense to form fine (<5 micron) solids. Removal of the particles/solids from the gas has been shown to be effective at reducing alkali content to required limits set by downstream equipment. Some work also has been done 11

12 on removal of alkali compounds at higher temperatures ( o C) using alkali getters, but additional R&D is needed on this approach. Stevens also addresses ammonia control. Ammonia forms in most biomass gasifiers from fuelbound nitrogen. Ammonia production is higher in pressurized gasifiers due to equilibrium considerations and in pyrolytic (indirectly-heated) gasifiers due to the stronger reducing environment. Ammonia removal is required for regulatory, but not generally for technical reasons. Removal will be required where there are strict NO x emissions regulations. For biomass conversion systems not located in ozone non-attainment areas, ammonia control is unlikely to be needed for regulatory reasons. Complete thermal decomposition of ammonia occurs only at >1000 o C, which is above temperatures typical for biomass gasification. At lower temperatures, ammonia has been shown to be destroyed (>90-99%) by dolomite, nickel-based steam reforming catalysts, alumino-silicate catalysts (Qadar et al., 1996), and other tar cracking catalysts at approximately the same temperatures required for tar cracking. Iron-based catalysts also work at somewhat higher temperature (900 o C). Such concepts have not been proven for large-scale, continuous operation. Wet scrubbing (with sulfuric acid) is well proven for ammonia removal at low temperatures. Removal of particles to some extent or other, which Stevens discusses at length, is essential for any biomass conversion system. There are a variety of commercially-established particle removal technologies available. Cyclones are a well-established technology for removing particles at any operating temperature relevant to biomass. Cyclones can remove >90% of particles larger than 5 microns with only a modest pressure drop. Partial removal of 1-5 micron particles is possible. Cyclones are not effective for removal of sub-micron particles. So-called barrier filters can remove particles down to 0.5 micron at elevated temperatures. For removal of smaller particles than this, the pressure drop through such filters is prohibitively high. Ceramic and metallic barrier filters have been tested. In limited-duration tests (170 hours) with hot gas derived from sugarcane bagasse by air-blown pressurized fluidized-bed gasification, ceramic candle filters showed good mechanical integrity. On the other hand ceramic filters have shown some susceptibility to reaction with alkali vapors. Metallic candle filters have successfully undergone longer-duration demonstration (at 350 o C) at the Varnamo biomass IGCC plant. Dry electrostatic precipitators (ESP) can operate at 500 o C or higher. Wet ESP operate only up to 65 o C. ESPs are best suited for large-scale facilities due to physical size and cost scale economies. Venturi wet scrubbers can remove >99.9% of particles larger than 2 microns and 95-99% of particles over 1 micron with pressure drop between 2.5 and 25 kpa. Sulfur removal is less of a concern with most biomass systems than with coal due to intrinsically low sulfur levels in raw biomass. For gas turbine power generation applications, removal of H 2 S from biomass-derived gas generally will not be required because the H 2 S will be oxidized to SO 2 in the gas turbine combustor, and the SO 2 concentrations in the exhaust gas will be well below regulated levels (due to intrinsically low sulfur content of biomass). For fuel cell use of biomass gas or for synthesis of many liquid fuels, H 2 S can be present only at sub-ppm levels, since the catalysts involved are poisoned by sulfur. Active removal of H 2 S from biomass gas will typically be required in these cases. Because most biomass conversion systems considered to date would not require H 2 S removal from the gas, sulfur removal is not addressed in any detail in the biomass conversion literature. Others, particularly in the coal field, have addressed H 2 S removal. Cicero et al. (2003) review both commercially-available sulfur removal technologies (typically operating at low temperatures -- < 100 o C), as well as promising advanced systems that 12

13 are still under development. These latter systems are generally aiming to remove sulfur at elevated temperatures ( o C). For process designs requiring H 2 S removal developed in this work, we utilize commercially established Rectisol technology, a low-temperature physical absorption processing that uses methanol as an absorbent. This is discussed further in the next section CO 2 Removal Bulk CO 2 removal from syngas streams is required in some of our process designs. We simulate CO 2 removal based on the performance of the commercial Rectisol process. A Rectisol process (operating in the range of 20 to 70 bar) can be designed for either separate removal of CO 2 and sulfur species from a gas stream or for co-removal of these species. Separate removal is required when the sulfur is to be converted to elemental sulfur for storage or sale. If co-removed with CO 2, H 2 S must either be oxidized (by processing with the CO 2 through a combustor, e.g., in the gas turbine power island) or be co-stored (with the CO 2 ) underground. (The low sulfur concentrations in raw syngas make feasible the oxidation route when H 2 S is co-removed with CO 2.) Rectisol plants with co-removal of CO 2 and sulfur species use a single absorber column. For separate removal of CO 2 and H 2 S, a two-column design is needed, with each species removed in a separate column and recovered separately. In either design, typically any CO 2 removal rate (up to 100%) can be achieved, and nearly all of the sulfur species can also be removed. Due to the small but finite solubilities of other syngas components (H 2, CO, and CH 4 ) in methanol, some quantities of these syngas components will be incidentally removed. In a one column design, one may expect a loss rate of 0.3% of the H 2, 1.5% of the CO, and 3% of the CH 4. (In a two column design, the loss rates would be about 50% higher, or 0.45%, 2.25%, and 4.5% respectively.) It is feasible, however, to recover these incidentally removed species at the Rectisol stripping column and recycle them, such that net loss rates are essentially zero. Some of our plant designs include CO 2 capture at two different places in the process. For such designs, some cost savings can be achieved by utilizing two absorbers with a common solvent regeneration column. (In the regenerator, the acid gases are stripped from the solvent by steam heating, and the solvent is recycled to the absorbers.) Solubilities of acid gases in methanol increase with decreasing temperature, so operating temperatures for good Rectisol performance are relatively low, necessitating a refrigeration plant. Aside from refrigeration, the only electricity consumption for a Rectisol plant comes from pumping solvent. (If incidental syngas components are recovered and recompressed for recycling, then there is additionally compressor power consumption.) Example Gas Conditioning Strategies Stevens (2001) describes the overall gas conditioning approaches used at several existing biomass-gasification demonstration facilities. For producing clean, cool ( o C) fuel gases he describes plants in the Netherlands, the UK, and the USA. The Amergas power station in Geertruidenberg, Holland, uses a Lurgi circulating fluidized-bed gasifier operating at o C on waste biomass materials and making gas that is co-fired with 13

14 coal in a boiler. This is currently the largest biomass gasification facility in the world (85 MW th gas output, or the equivalent of approximately 700 tonnes per day dry biomass input). The gas conditioning system consists of cyclones for particulate removal, followed by gas cooling (by steam raising) to o C, followed by bag (fabric) filters for additional particulate removal, followed by wet scrubbers for tar removal and gas cooling. The tars are collected and injected into the gasifier. There is an additional final wet scrubbing for ammonia removal, with the ammonia being re-injected into the gasifier for thermal destruction and reaction with NO x. The UK project, called ARBRE, is a biomass-gasifier IGCC (8 MW e ) that is intended to operate on short-rotation woody biomass fuel grown by farmers in the area (Pitcher et al., 1998; Maniatis, 2001). Plant construction was completed and commissioning started in 2002, but the plant is not yet operating due to ownership uncertainties. The gasifier is an air-blown, atmospheric-pressure circulating fluidized-bed designed by the Swedish company, TPS. After bed solids are captured (via cyclone) for recycle, the raw product gas passes to a hightemperature fluidized-bed tar cracker that uses dolomite as a catalyst. The cracker is followed by a syngas cooler generating steam. The warm gas (350 o C) then passes through fabric filters for capture of particles and condensed alkali. Finally, wet scrubbing with dilute H 2 SO 4 (sulfuric acid) is used as the final conditioning to remove particles, residual tars, and ammonia. The USA project is located in Burlington, Vermont. It consists of a FERCO/Battelle indirectlyheated (pyrolytic) gasifier with a capacity to gasify 200 dry tonnes per day of forest residues. The pyrolytic gases leave the gasifier at about 800 o C and are passed through a cyclone for bulk particle removal and then sent at elevated temperature to be burned in an existing boiler. The plan (which has not yet been put into effect) is to install a catalytic tar cracker following the cyclone, pass the cracked gas through a syngas cooler to raise some steam, and then use a fabric filter and wet scrubbing to remove remaining particles, tars, and ammonia. The proposed catalyst for tar cracking is DN34, mentioned earlier. Stevens also describes the overall gas conditioning strategy for producing warm ( o C) or hot (> 500 o C) clean fuel gases. He cites the Varnamo project in Sweden where this strategy was pursued with success. The Varnamo facility included a pressurized air-blown gasifier with limestone or dolomite bed material to provide partial tar cracking. The resulting tar level in the gas was about 5 g/nm 3. By insulating downstream pipes/filters to keep the gas temperature above ~350 o C, tar condensation was prevented. Bulk particulate removal was effected with a primary cyclone, which was followed by a syngas cooler generating steam and dropping the gas temperature to about 350 o C. A barrier filter operating at this temperature provided the final contaminant removal step. Ceramic candle filters were initially tested, but subsequently metallic candle filters were found to have more satisfactory performance Gas Conditioning for Ultra Clean Applications For fuel cell and liquid fuels synthesis applications (e.g., production of dimethyl ether or Fischer- Tropsch liquids), gas contaminant levels, including sulfur, generally must be considerably lower than for gas turbine applications. There is considerably less analysis or empirical results published on biomass-related gas conditioning issues for such ultra clean applications. Dayton (2001) reviews gas quality requirements relating to particulates, sulfur, ammonia, halogens (e.g., hydrochloric acid), and alkali metals for different fuel cell designs (proton exchange membrane, phosphoric acid, molten carbonate, and solid oxide). For synthesis of liquid fuels, Graham and Bain (1993), as cited by Stevens (2001), define gas quality requirements for syngas to be used 14

15 for fuels synthesis. Nitrogen-dilution of the gas is unacceptable due primarily to the increased costs associated with the higher volumetric flow of gas that would need to be processed. Pressurized gasification is preferred since synthesis processes are favored by higher pressures. Particle loading in the clean synthesis gas must be <0.02 mg/nm 3. Tar loading in the clean gas must be < 0.1 mg/nm 3. Sulfur loading in the syngas must be < 0.1 mg/nm Conversion of Synthesis Gas to Final Products There is considerable commercial activity relating to conversion of synthesis gas to liquid fuels and to hydrogen. With a few important exceptions (discussed below), in most such commercial facilities, the synthesis gas is produced from natural gas. The conversion of clean synthesis gas to liquid fuels, including F-T fuels and DME, involves passing the gas over catalysts that promote the desired synthesis reactions. These reactions are exothermic, and the reactor temperature will increase as the reactions proceed if no heat is removed. Higher temperatures promote faster reactions, but maximum conversion is favored by lower temperatures. Also, catalysts are deactivated when overheated. Thus, the temperature rise in a synthesis reactor must be controlled. In practice, a reactor operating temperature of o C balances kinetic, equilibrium, and catalyst activity considerations. Two available synthesis reactor designs, gas-phase (or fixed-bed) and liquid-phase (or slurryreactor), handle temperature control using different approaches. The basic gas-phase design involves the flow of syngas over a fixed-bed of catalyst pellets. The basic liquid-phase design involves bubbling syngas through an inert mineral oil containing powdered catalyst in suspension. In a gas-phase reactor it is difficult to maintain isothermal conditions by direct heat exchange (due to low gas-phase heat transfer coefficients). To limit temperature rise, the synthesis reactions are staged, with cooling between reactor stages. Also, by limiting the initial concentration of CO entering the reactor (to vol%) the extent of the exothermic reactions can be controlled. Control of the CO fraction is achieved in practice by maintaining a sufficiently high recycle of unconverted H 2 -rich syngas back to the reactor. In a liquid phase reactor, the feed gas is bubbled through an inert oil in which catalyst particles are suspended (Figure 6). Much higher heat release rates (i.e., extents of reaction) can be accommodated without excessive temperature rise as compared to a gas-phase reactor because of more effective reactor cooling by boiler tubes immersed in the fluid. The vigorous mixing, intimate gas-catalyst contact, and uniform temperature distribution enable a high conversion of feed gas to liquids in a relatively small reactor volume. Conversion by liquid-phase F-T synthesis is especially high achieving a single-pass fractional conversion of CO to F-T liquids of about 80% (Bechtel Group, 1990). This compares to less than 40% conversion with traditional fixed-bed F-T reactors. High single-pass conversion rates make liquid-phase reactors especially attractive from an overall cost perspective for co-production process designs (Larson and Ren, 2003), whereby the synthesis gas is passed once through the synthesis reactor, and any unconverted gas is used to generate co-product electricity for sale (e.g., Figure 3, upper). Liquid-phase reactors are commercially available for F-T synthesis. They have been demonstrated at pilot scale for DME 15

16 synthesis (and at commercial scale for methanol synthesis, a catalytic process closely related to DME synthesis). In our work, we have examined both once-through and recycle synthesis configurations for producing DME, but given the very high single-pass conversion achievable with F-T synthesis, we have chosen to consider only once-through configurations for F-T synthesis. Fuel product (vapor) + unreacted syngas Steam Disengagement zone Catalyst powder slurried in oil TYPICAL REACTION CONDITIONS: P = atmospheres T = o C Cooling water CO catalyst CH 3 OCH 3 CH 3 OH Synthesis gas (CO + H 2 ) H 2 C n H 2n+2 (depending on catalyst) Figure 6. Simplified depiction of a liquid-phase synthesis reactors, which yield higher one-pass conversion of (CO+H 2 ) to liquids than traditional gas-phase reactors Synthesis of Fischer-Tropsch Fuels Fischer-Tropsch conversion was first done commercially, based on gas-phase technology, in the 1930s when Germany started producing F-T liquids from coal syngas as vehicle fuel (Dry, 2002). A coal-to-fuels program has been operating in South Africa since the early 1950s. Starting in the 1990s, there has been renewed interest globally in F-T synthesis to produce liquids from large reserves of remote stranded natural gas that has little or no value because of its distance from markets (Oukaci, 2002; Rahmim, 2003). Of particular interest is the production of middle distillate fuels (diesel-like fuels) with unusually high cetane numbers and containing little or no sulfur or aromatics. Such fuels (derived by natural gas F-T conversion) are now beginning to be blended with conventional diesel fuels, e.g., in California, to meet increasingly strict vehicle fuel specifications designed to reduce tailpipe emissions. In addition to Shell s well-known gas-to-liquids (GTL, used synonymously with gas-to-ft liquids) plant (using gas-phase synthesis technology) in Malaysia that has been operating for nearly a decade, there are additional large commercial GTL facilities under construction or at advanced planning stages, including: 34,000 barrels/day plant, Qatar Petroleum project that will use Sasol liquid-phase synthesis technology; slated to come on line in ,000 b/d, Chevron project in Nigeria, also using Sasol FT technology; to come on line by

17 140,000 b/d, Shell project in Qatar, using Shell s gas-phase FT technology; to come on line in ,000 b/d ExxonMobil project in Qatar, using ExxonMobil liquid-phase technology; to come on line in 2011 There is also a growing resurgence of interest in production of F-T fuels from gasified coal. Coal-based FT fuel production was commercialized beginning with the Sasol I, II, and III plants (175,000 b/d total capacity) built between 1956 and 1982 in South Africa. (Sasol I is now retired). In the U.S., a major assessment was recently completed for the U.S. Department of Energy on co-production of F-T liquids and electricity from coal via gasification at large scales (Bechtel et al. 2003a; 2003b), and the U.S. Department of Energy is cost-sharing a $0.6 billion demonstration project in Gilberton, Pennsylvania, that will make from coal wastes 5,000 b/d of F-T liquids using liquid-phase synthesis and 41 MW e of electricity. Also, there are proposals for 33,000 b/d and 57,000 b/d facilities for FT fuels production from coal in the state of Wyoming. China s first commercial coal-ft project is under construction by the Shenhua Group in Inner Mongolia. The plant is slated to produce 1 million tonnes of oil products/year (20,000 barrels/day or 0.3 billion gallons per year) when it comes on line in China has also signed a letter of intent with Sasol for two coal-ft plants that will produce together 6 million tonnes of liquid fuels per year (120,000 b/d). Prefeasibility analysis is ongoing for these plants. The projected capital investment required for these plants is a total of $6 billion. Meanwhile, Chinese researchers have been developing pilot-scale slurry phase FT synthesis technology for coalderived syngas (e.g., Li et al., 2001; Yang et al., 2003). The process for converting biomass into F-T liquids is similar in many respects to that for converting coal. Preliminary technical/economic analyses on biomass conversion was published by Larson and Jin (1999a and 1999b). More recently, there have been several detailed technical and economic assessments published (Bechtel, 1998; Tijmensen, 2000; Tijmensen et al., 2002; Hamelinck et al., 2003; Boerrigter and van der Drift, 2003; Hamelinck et al., 2004). There is considerable current interest in Europe in production of FT fuels from biomass, motivated by large financial incentives. For example, in the UK a 20 pence per liter ($1.40/gal) incentive for biomass-derived diesel fuel has been in place since July Incentives are also in place in Germany, Spain, and Sweden. Such incentives have been introduced in part as a result of European Union Directive 2003/30/EC, which requires all member states to have 2% of all petrol and diesel consumption (on an energy basis) be from biofuels or other renewable fuels by the end of 2005, reaching 5.75% by the end of Synthesis of Dimethyl Ether The leading commercial developer of fixed-bed DME synthesis reactor designs is Halder- Topsoe. 6 Mobil and Snamprogetti S.p.A. hold patents for DME synthesis processes (Zahner, 1977; Pagani, 1978), but at present are not pursuing commercial development of the technology. Leading private developers of slurry-bed DME synthesis reactors are Air Products and Chemicals, Inc. (APCI) (Lewnard et al., 1990; Brown et al., 1991; Lewnard et al., 1993; Air Products and Chemicals, 1993; Peng et al., 1999a; Peng et al., 1999b) and the NKK Corporation 6 The fixed-bed design of Halder-Topsoe includes three stages of synthesis reactors with cooling between each stage and recycle of unconverted syngas (Hansen et al., 1995). The patent for this process specifies a feed gas CO concentration of less than 10% and a recycle volume of unconverted syngas ranging from 93% to 98% of the total unconverted syngas (Voss et al., 1999). The fraction of CO converted on a single pass through each reactor stage (assuming a three-stage intercooled set of reactors) ranges from 16% to 34%, depending on the H 2 /CO ratio. 17